Great discussion, very thoughtful back and forth. At 51:20, I was happy to hear Reuben say, in regard to level, "...does somebody push something beyond where it musically feels like it wants to be. You've got to let the music guide you."
This is a smart guy and really well spoken. A very interesting discussion. Thanks. IMHO, discussions of sample rate should always start with the Nyquist Theorem - which holds that there is no advantage (as in none, zero) to rates over twice the highest frequency you want to capture. This is counter-intuitive (1). It took me a long time to come to terms with this, because it seems that more samples would equal higher "resolution." For me (and I'm always wrong), the breakthrough came when I realized I was thinking of waves sideways-on, rather than coming straight at me, as they do from a sound source. The following was helpful: picture yourself in a chair reading a book at the beach with waves coming directly at you. How often would you have to look up from your book (that's the sample rate as it turns out) to not miss seeing a wave (ie have one escape your detection) even at the shortest time period they might follow each other? Turns out the answer (how often would you have to look up) seems to be "twice the highest rate," or better, "half that time period" (between wave peaks). (1) We think more samples will give us a better "picture" of a wave, but that's the job of the amplitude measurement. More bits can better capture changes in amplitude because more bits can write down finer changes in numbers. The sample rate only needs to reliably capture the "frequency" of the peaks. Once we have sufficient frequency and amplitude data, we can reconstruct the analog wave.
Excellent video, I really appreciate the in-depth knowledge of mastering and it was priceless, I learned a lot. Thank you!
Great discussion, very thoughtful back and forth. At 51:20, I was happy to hear Reuben say, in regard to level, "...does somebody push something beyond where it musically feels like it wants to be. You've got to let the music guide you."
This is a smart guy and really well spoken. A very interesting discussion. Thanks.
IMHO, discussions of sample rate should always start with the Nyquist Theorem - which holds that there is no advantage (as in none, zero) to rates over twice the highest frequency you want to capture. This is counter-intuitive (1). It took me a long time to come to terms with this, because it seems that more samples would equal higher "resolution."
For me (and I'm always wrong), the breakthrough came when I realized I was thinking of waves sideways-on, rather than coming straight at me, as they do from a sound source. The following was helpful: picture yourself in a chair reading a book at the beach with waves coming directly at you. How often would you have to look up from your book (that's the sample rate as it turns out) to not miss seeing a wave (ie have one escape your detection) even at the shortest time period they might follow each other? Turns out the answer (how often would you have to look up) seems to be "twice the highest rate," or better, "half that time period" (between wave peaks).
(1) We think more samples will give us a better "picture" of a wave, but that's the job of the amplitude measurement. More bits can better capture changes in amplitude because more bits can write down finer changes in numbers. The sample rate only needs to reliably capture the "frequency" of the peaks. Once we have sufficient frequency and amplitude data, we can reconstruct the analog wave.
this is gold